The Best Basketball Player of our Generation

By: Branden Tsang

The 2020-2021 NBA season is coming to a close as the Milwauke Bucks are currently facing off against the Phoenix Suns in a best of 7 series. This past season was a wild return from the NBA bubble which forced players to stay confined in a resort at Disney World near Orlando Florida the season prior. This season had a tremendous amount of injuries including major figures such as Lebron James, James Harden, Kyrie Irving, Kawhi Leonard, Jamal Murray and much more.

As we all know, basketball is a team sport where 5 players on one team play against 5 players on an opposing team. Each player has stats such as points, assists, rebounds, etc. which contributes to how great a player is. Decades ago, Michael Jordan shocked the world with his jaw dropping performances and scoring prowess. Following him Kobe Bryant carried the torch and was dubbed as the next Michael Jordan. Later came Lebron James with his abnormal size, speed, and longevity.

While both Michael Jordan and Kobe have retired, Lebron James is still playing and proving his case for being the greatest player. Rivaling him currently are big name players such as: Stephen Curry a 2 time MVP and 3 time NBA champion; Kevin Durant, a NBA MVP and 2 time NBA champion. Kawhi Leonard, James Harden, and Anthony Davis are all more names that are amazing players with great stats. These players's success story all begs the question: Who is the greatest player of our generation?

image.png

Data Gathering
To begin we will import data from an excel file which contains data from players drafted from 1996 to the 2019 NBA season. This data is not up to date with the current NBA season but will be considering players up to the 2019 NBA season. Our data contains information pertaining to the amount of games played, points per/game, field goal percentage, rebounds, assists etc. Once we have this, we can begin tidying our data and figure out which player was the best player.

Data Tidying
Once we have all our data, we can begin sifting through and removing players who were drafted yet did not play in the NBA and also remove players who barely have any games played which I will deem as 30 or less games (a little less than half a season). This will help remove small sample data that could skew the results we are hoping to get. For example, someone who played 2 minutes and made a single shot will have a 100% field goal percentage which I hope does not give him credentials to be considered the greatest player for having never missed a shot in his career

The next step of our data tidying is to first find the best shooter of our generation and remove data that we don't find relevant. Being a good player means that you must be good at scoring the basket which is key to winning games. We can also create a new dataset that has only the information we need to conduct this research such as field goal percentage, points, games, etc.

Visualization
Now I plotted the data on a scatterplot seeing how shooting has evolved from 1996 as well as shooting data compared to games played. The first graph will be used to check whether or not shooting in recent years is better than shooting in previous years as the game develops and more players have access to the previous great players' game replays. The second graph is used to check whether players who played more games truly do shoot better than players who play less. Logically speaking, players who play more games should have a higher shooting percentage than those who play less due to the fact that they either practice more, or simply play better than other players therefore they are given a bigger role.

More Visualization
To get a better understanding of where each player stands compared to others I decided to create a formula for each player and added their offensive stats to their defensive stats and their points in order to denote who had the most impact on both ends of the floor.

More Visualization and Machine Learning
For more visualization, I decided to plot the score of the teams on a map of the US denoted by where their stadiums are located. If they had an above average score, they would be plotted in green and below average will be red. The deeper the hue indicates the further they are from the average.
To get the location of the stadiums, we scraped data from the web in order to plot the latitude and logitude of the stadiums on a map.

Now create a dataframe for scores for each team by adding up all the total players' scores that played for that team each season. This will help determine which team was the best throughout the past 25 years. We also wish to add championship data information in order to determine which team was the best team.

Now with the dataframe with the stadium data as well as the score, we can get data on the residuals from our model using statsmodels.

Here we plot a pie chart of the data of how each team does relative to others in terms of the team's total score from 1996-2019. From the looks of it, the Los Angeles Lakers should have the highest percentage of the pie chart followed by the Houston Rockets. The New Orleans Pelicans have the lowest percentage of the total score of being a small market team and having few super stars in the past 25 years.

Here we plot the pie chart of championships won by teams. Basketball is a team sport where a team's success is more important than individual player success. In recent NBA history the Lakers have won a majority of the championships having won 5 from 2000-2010.

Now we can plot our data on a map, marking each of the teams' stadiums. Teams that had worse overall performers will be red colored, whereas teams that perform better will slowly change colors to become more green. The more green colored icons will be teams that performed better. When clicking on the icon, the team as well as their overall score will popup.

From the map we can see the Los Angeles Lakers were the most dominant team in the last 25 years being the only team that has a bold green color. Most of the teams in the league were a light green color. The worst teams were the NOP and OKC having been founded in 2002 and 2008 respectively which gives them less history and less players who were noteworthy during this time.

Conclusion
The NBA is a very competitive league in which players grind day in and day out. Players are always striving to get better at scoring, rebounding, assisting and teams are always striving to get more wins. A team's success is driven by their players' success and a players' success is driven by a team's success. In the end, the coveted prize in the NBA championship.
Look at James Harden and Kobe Bryant for example, both players with amazing scoring abilities. One player played a majority of his career in Houston for the Rockets while the other played as a lifetime Los Angeles Laker. One player has 3 of the top 5 spots in our scoring list and the other doesn't. However, the player with the top individual spots has accomplished essentially nothing for his organization and won 0 championships whereas the other has won 5 NBA championships.
Individual talent is needed to prove your worth as a great player. However all great players must also have a great basketball resume to show that they are champions. Basketball is a team sport where 5 players play against an opposing team of 5 players. In our generation, Kobe Bryant has shown what it takes to become a champion. He showed that he can lead a team to the promise land and deliver wins to his city and team on the biggest stage while being the best and most dominant basketball savant we have seen since the days of Michael Jordan.